- Kernel - mailweb.openeuler.org

[PATCH OLK-6.6 v2] config: Open CONFIG_AARCH32_EL0 and close CONFIG_ARM64_ILP32
by Jinjie Ruan 08 Dec '23

08 Dec '23

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8JVJ3 CVE: NA ------------------------------------------------------------------------- No one currently uses IPL32, close it to simplify the code logic of TASK_SIZE_MAX, thereby improving benchmark performance. For example, all test items of libMicro can be improved by 4.89% on average. Since commit 7f6675f0e994 ("arm64: rename COMPAT to AARCH32_EL0"), CONFIG_COMPAT has been renamed to CONFIG_AARCH32_EL0, so open CONFIG_AARCH32_EL0 as what CONFIG_COMPAT used to be. Signed-off-by: Zhen Lei <thunder.leizhen(a)huawei.com> Reviewed-by: Cheng Jian <cj.chengjian(a)huawei.com> Reviewed-by: Liu Chao (CR) <liuchao173(a)huawei.com> Reviewed-by: Xiongfeng Wang <wangxiongfeng2(a)huawei.com> Signed-off-by: Zheng Zengkai <zhengzengkai(a)huawei.com> Signed-off-by: Jinjie Ruan <ruanjinjie(a)huawei.com> --- arch/arm64/configs/openeuler_defconfig | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/arch/arm64/configs/openeuler_defconfig b/arch/arm64/configs/openeuler_defconfig index 9d2f717c1f7c..ce5a7aeaeeca 100644 --- a/arch/arm64/configs/openeuler_defconfig +++ b/arch/arm64/configs/openeuler_defconfig @@ -449,9 +449,10 @@ CONFIG_MITIGATE_SPECTRE_BRANCH_HISTORY=y # CONFIG_RODATA_FULL_DEFAULT_ENABLED is not set # CONFIG_ARM64_SW_TTBR0_PAN is not set CONFIG_ARM64_TAGGED_ADDR_ABI=y -CONFIG_COMPAT=y +CONFIG_AARCH32_EL0=y # CONFIG_KUSER_HELPERS is not set # CONFIG_COMPAT_ALIGNMENT_FIXUPS is not set +# CONFIG_COMPAT_TASK_SIZE is not set CONFIG_ARMV8_DEPRECATED=y CONFIG_SWP_EMULATION=y CONFIG_CP15_BARRIER_EMULATION=y @@ -526,6 +527,8 @@ CONFIG_EFI=y CONFIG_DMI=y # end of Boot options +CONFIG_COMPAT=y + # # Power management options # -- 2.34.1

2 1

[PATCH OLK-6.6 0/3] improve gettimeofday() performance in user space
by Yang Yingliang 08 Dec '23

08 Dec '23

Don't trap kernel to read the counter register, do this in userspace to improve performance. Yang Yingliang (3): arm64: arch_timer: Disable CNTVCT_EL0 trap if workaround is enabled vdso: do cntvct workaround in the VDSO arm64: arch_timer: disable CONFIG_ARM_ARCH_TIMER_WORKAROUND_IN_USERSPACE arch/arm64/configs/openeuler_defconfig | 1 + arch/arm64/include/asm/vdso/gettimeofday.h | 32 ++++++++++++++++++++++ drivers/clocksource/Kconfig | 5 ++++ drivers/clocksource/arm_arch_timer.c | 27 ++++++++++++++++++ include/linux/clocksource.h | 4 +++ include/vdso/datapage.h | 8 +++++- kernel/time/vsyscall.c | 16 +++++++++++ 7 files changed, 92 insertions(+), 1 deletion(-) -- 2.25.1

2 4

[PATCH OLK-5.10] locking/osq_lock: Avoid false sharing in optimistic_spin_node
by Zeng Heng 08 Dec '23

08 Dec '23

hulk inclusion category: performance bugzilla: https://gitee.com/openeuler/kernel/issues/I8MV01 -------------------------------- In the prologue of osq_lock(), it set `cpu` member of percpu struct optimistic_spin_node with the local cpu id, after that the value is never changed any more. Consequently, we can regard the `cpu` member as a constant variable. In the meanwhile, other members like next, prev and locked are frequently modified by osq_lock() and osq_unlock() which come from rwsem, mutex and so on. So here we can place padding and locate them at different cache lines, so that avoid cache misses when access optimistic_spin_node.cpu field by other CPUs from the node list. Here provide the UnixBench full-core test result as below: Machine Intel(R) Xeon(R) Gold 6248 CPU, 40 cores, 80 threads Run the command of "./Run -c 80 -i 3" 10 times and take the average. System Benchmarks Index Values Without Patch With Patch Diff Dhrystone 2 using register variables 185876.43 185945.41 0.04% Double-Precision Whetstone 79637.27 79659.29 0.03% Execl Throughput 9909.61 10576.06 6.73% File Copy 1024 bufsize 2000 maxblocks 1723.01 2086.08 21.07% File Copy 256 bufsize 500 maxblocks 1150.24 1338.21 16.34% File Copy 4096 bufsize 8000 maxblocks 3719.19 4011.99 7.87% Pipe Throughput 66184.84 66025.25 -0.24% Pipe-based Context Switching 30606.18 31074.21 1.53% Process Creation 9442.48 9450.77 0.09% Shell Scripts (1 concurrent) 44526.52 46548.54 4.54% Shell Scripts (8 concurrent) 42903.96 45718.56 6.56% System Call Overhead 3645.20 3717.42 1.98% ======== System Benchmarks Index Score 15126.87 15931.29 5.32% Signed-off-by: Zeng Heng <zengheng4(a)huawei.com> --- include/linux/osq_lock.h | 2 +- kernel/locking/osq_lock.c | 3 ++- 2 files changed, 3 insertions(+), 2 deletions(-) diff --git a/include/linux/osq_lock.h b/include/linux/osq_lock.h index 5581dbd3bd34..deb90ad5f560 100644 --- a/include/linux/osq_lock.h +++ b/include/linux/osq_lock.h @@ -9,7 +9,7 @@ struct optimistic_spin_node { struct optimistic_spin_node *next, *prev; int locked; /* 1 if lock acquired */ - int cpu; /* encoded CPU # + 1 value */ + int cpu ____cacheline_aligned; /* encoded CPU # + 1 value */ }; struct optimistic_spin_queue { diff --git a/kernel/locking/osq_lock.c b/kernel/locking/osq_lock.c index 1de006ed3aa8..f7e7b76f6274 100644 --- a/kernel/locking/osq_lock.c +++ b/kernel/locking/osq_lock.c @@ -96,7 +96,8 @@ bool osq_lock(struct optimistic_spin_queue *lock) node->locked = 0; node->next = NULL; - node->cpu = curr; + if (!node->cpu) + node->cpu = curr; /* * We need both ACQUIRE (pairs with corresponding RELEASE in -- 2.25.1

2 1

[PATCH OLK-6.6 v3 00/19] Add sharepool support v3
by Wang Wensheng 08 Dec '23

08 Dec '23

This patch series migrates share pool features from OLK-5.10 to OLK-6.6. We don't want to do this work patch by patch since there are hundreds of patches in OLK-5.10 including bugfix and small features generation. Instead we just take the final version and split it into a few small patches, each of which contains one feature at the final version. The share pool features is a big feature, it is mainly used to share user virtual memory for different processes in the same group. It could be used by this steps: Process A create a new group which is owned by process A. Process A add process B to the group. Process A add process C to the same group. Process B alloc a new memory VA, and write something in it. The VA was send to the process C by IPC, then process C got it. The process C access the VA and got the data directly. The process A could add more processes in the group to share the memory. Fix the memory by use the free function or exit the group. The new features is enabled both by CONFIG_SHARE_POOL and the enable_ascend_share_pool bootarg, it would not affect anything if disabled. v1->v2: Delete some unused code. v2->v3: Rename several function. Wang Wensheng (19): mm/hugetlb: Introduce hugetlb_insert_hugepage_pte[_by_pa] mm/vmalloc: Extend vmalloc usage about hugepage mm: Extend mmap assocated functions to accept mm_struct mm/sharepool: Add base framework for share_pool mm/sharepool: Add sp_area management code mm/sharepool: Reserve the va space for share_pool mm/sharepool: Implement mg_sp_make_share_u2k() mm/sharepool: Implement mg_sp_unshare_kva mm/sharepool: Implement mg_sp_walk_page_range() mm/sharepool: Implement mg_sp_free() mm/sharepool: Implement mg_sp_alloc() mm/sharepool: Implement mg_sp_make_share_k2u() mm/sharepool: Implement mg_sp_group_add_task() mm/sharepool: Implement mg_sp_group_id_by_pid() mm/sharepool: Implement mg_sp_id_of_current() mm/sharepool: Implement mg_sp_config_dvpp_range() mm/sharepool: Add proc interfaces to show sp info mm/sharepool: support fork() and exit() to handle the mm mm/sharepool: Protect the va reserved for sharepool include/linux/hugetlb.h | 32 + include/linux/mempolicy.h | 11 + include/linux/mm.h | 17 + include/linux/mm_types.h | 6 + include/linux/share_pool.h | 267 +++ include/linux/vmalloc.h | 18 + kernel/fork.c | 3 + mm/Kconfig | 34 + mm/Makefile | 1 + mm/gup.c | 23 +- mm/hugetlb.c | 80 + mm/mempolicy.c | 12 +- mm/mmap.c | 50 +- mm/mremap.c | 4 + mm/share_pool.c | 3448 ++++++++++++++++++++++++++++++++++++ mm/util.c | 4 + mm/vmalloc.c | 205 ++- 17 files changed, 4194 insertions(+), 21 deletions(-) create mode 100644 include/linux/share_pool.h create mode 100644 mm/share_pool.c -- 2.17.1

2 20

[PATCH OLK-5.10] roh/core: Synchronously update the mac address of
by Ke Chen 08 Dec '23

08 Dec '23

driver inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8MTN4 ----------------------------------------------------------------------- In ROH mode, when the IP address of a VLAN device changes, we need to change the MAC address of the VLAN device to maintain the mapping between the IP address, EID, and MAC address. In this way, the ARP fast response in ROH mode can take effect. Fixes:55901e4aae1e("roh/core: Add support for inetaddr notifier in roh/core") Signed-off-by: Ke Chen <chenke54(a)huawei.com> Reviewed-by: Gang Zhang <gang.zhang(a)huawei.com> --- drivers/roh/core/core.c | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) diff --git a/drivers/roh/core/core.c b/drivers/roh/core/core.c index 1b8041004be3..9b867f2cc2d5 100644 --- a/drivers/roh/core/core.c +++ b/drivers/roh/core/core.c @@ -3,6 +3,7 @@ #include <linux/pci.h> #include <linux/inetdevice.h> +#include <linux/if_vlan.h> #include "core_priv.h" #include "core.h" @@ -245,12 +246,26 @@ static int roh_ipv4_event(struct notifier_block *this, unsigned long event, void struct in_ifaddr *ifa = ptr; struct roh_device *device; struct net_device *ndev; + struct sockaddr s_addr; struct sockaddr_in in; int ret; + if (event != NETDEV_UP) + return NOTIFY_DONE; + device = container_of(this, struct roh_device, nb); ndev = ifa->ifa_dev->dev; - if (device->netdev != ndev || event != NETDEV_UP) + + if (is_vlan_dev(ndev)) { + if (vlan_dev_real_dev(ndev) == device->netdev) { + s_addr.sa_family = ndev->type; + u64_to_ether_addr(be32_to_cpu(ifa->ifa_address) & 0xffffff, s_addr.sa_data); + dev_set_mac_address(ndev, &s_addr, NULL); + } + return NOTIFY_DONE; + } + + if (device->netdev != ndev) return NOTIFY_DONE; in.sin_addr.s_addr = ifa->ifa_address; -- 2.30.0

2 1

[PATCH OLK-6.6 1/1] ARM: Add unwinding annotations to __loop.*delay functions
by Xu Kuohai 08 Dec '23

08 Dec '23

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8MTBB CVE: NA -------------------------------- The unwind directives can't handle nested function entries, so only the outermost function __loop_udelay is annotated, and the inner nested entries share the same unwind info with __loop_udelay. Signed-off-by: Xu Kuohai <xukuohai(a)huawei.com> Reviewed-by: Yang Jihong <yangjihong1(a)huawei.com> Signed-off-by: Zheng Zengkai <zhengzengkai(a)huawei.com> --- arch/arm/lib/delay-loop.S | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/arch/arm/lib/delay-loop.S b/arch/arm/lib/delay-loop.S index 3ac05177d097..3ba37ce6dbb0 100644 --- a/arch/arm/lib/delay-loop.S +++ b/arch/arm/lib/delay-loop.S @@ -7,6 +7,7 @@ #include <linux/linkage.h> #include <asm/assembler.h> #include <asm/delay.h> +#include <asm/unwind.h> #ifdef CONFIG_ARCH_RPC .arch armv4 @@ -25,6 +26,7 @@ */ ENTRY(__loop_udelay) +UNWIND(.fnstart) ldr r2, .LC1 mul r0, r2, r0 @ r0 = delay_us * UDELAY_MULT ENTRY(__loop_const_udelay) @ 0 <= r0 <= 0xfffffaf0 @@ -58,6 +60,7 @@ ENTRY(__loop_delay) #endif bhi __loop_delay ret lr -ENDPROC(__loop_udelay) -ENDPROC(__loop_const_udelay) ENDPROC(__loop_delay) +ENDPROC(__loop_const_udelay) +UNWIND(.fnend) +ENDPROC(__loop_udelay) -- 2.30.2

2 2

[PATCH OLK-6.6 1/1] ARM: Add unwinding annotations to __loop.*delay functions
by Xu Kuohai 08 Dec '23

08 Dec '23

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I8MTBB CVE: NA -------------------------------- The unwind directives can't handle nested function entries, so only the outermost function __loop_udelay is annotated, and the inner nested entries share the same unwind info with __loop_udelay. Signed-off-by: Xu Kuohai <xukuohai(a)huawei.com> Reviewed-by: Yang Jihong <yangjihong1(a)huawei.com> Signed-off-by: Zheng Zengkai <zhengzengkai(a)huawei.com> --- arch/arm/lib/delay-loop.S | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/arch/arm/lib/delay-loop.S b/arch/arm/lib/delay-loop.S index 3ac05177d097..3ba37ce6dbb0 100644 --- a/arch/arm/lib/delay-loop.S +++ b/arch/arm/lib/delay-loop.S @@ -7,6 +7,7 @@ #include <linux/linkage.h> #include <asm/assembler.h> #include <asm/delay.h> +#include <asm/unwind.h> #ifdef CONFIG_ARCH_RPC .arch armv4 @@ -25,6 +26,7 @@ */ ENTRY(__loop_udelay) +UNWIND(.fnstart) ldr r2, .LC1 mul r0, r2, r0 @ r0 = delay_us * UDELAY_MULT ENTRY(__loop_const_udelay) @ 0 <= r0 <= 0xfffffaf0 @@ -58,6 +60,7 @@ ENTRY(__loop_delay) #endif bhi __loop_delay ret lr -ENDPROC(__loop_udelay) -ENDPROC(__loop_const_udelay) ENDPROC(__loop_delay) +ENDPROC(__loop_const_udelay) +UNWIND(.fnend) +ENDPROC(__loop_udelay) -- 2.30.2

2 1

[PATCH 0/5] LoongArch: add cpufreq and ls2k500 bmc support
by Hongchen Zhang 08 Dec '23

08 Dec '23

Chong Qiao (4): fbdev: add ls2k500sfb driver for ls2k500 bmc. ipmi: add ls2k500 bmc ipmi support. LoongArch: defconfig: enable CONFIG_FB_LS2K500=m. LoongArch: fix ls2k500 bmc not work when installing iso liuyun (1): cpufreq: Add cpufreq driver for LoongArch arch/loongarch/Kconfig | 1 + arch/loongarch/configs/loongson3_defconfig | 5 + arch/loongarch/include/asm/fpu.h | 13 +- drivers/char/ipmi/Makefile | 4 + drivers/char/ipmi/btlock.h | 92 ++ drivers/char/ipmi/ipmi_si.h | 11 + drivers/char/ipmi/ipmi_si_intf.c | 4 + drivers/char/ipmi/ipmi_si_ls2k500.c | 173 +++ drivers/char/ipmi/kcs_bmc_ls2k500.h | 67 + drivers/cpufreq/Kconfig | 11 + drivers/cpufreq/Makefile | 1 + drivers/cpufreq/loongson3-acpi-cpufreq.c | 1527 ++++++++++++++++++++ drivers/gpu/drm/loongson/loongson_module.c | 15 + drivers/video/fbdev/Kconfig | 13 + drivers/video/fbdev/Makefile | 1 + drivers/video/fbdev/ls2k500sfb.c | 792 ++++++++++ 16 files changed, 2729 insertions(+), 1 deletion(-) create mode 100644 drivers/char/ipmi/btlock.h create mode 100644 drivers/char/ipmi/ipmi_si_ls2k500.c create mode 100644 drivers/char/ipmi/kcs_bmc_ls2k500.h create mode 100644 drivers/cpufreq/loongson3-acpi-cpufreq.c create mode 100644 drivers/video/fbdev/ls2k500sfb.c -- 2.33.0

1 5

[PATCH OLK-5.10] module: Expose module_init_layout_section()
by Zhao Wenhui 08 Dec '23

08 Dec '23

From: James Morse <james.morse(a)arm.com> stable inclusion from stable-v5.10.194 commit ecd62c85120dc74b07c5ccd1532de121e81a4dba category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I83LMD Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- commit 2abcc4b5a64a65a2d2287ba0be5c2871c1552416 upstream. module_init_layout_section() choses whether the core module loader considers a section as init or not. This affects the placement of the exit section when module unloading is disabled. This code will never run, so it can be free()d once the module has been initialised. arm and arm64 need to count the number of PLTs they need before applying relocations based on the section name. The init PLTs are stored separately so they can be free()d. arm and arm64 both use within_module_init() to decide which list of PLTs to use when applying the relocation. Because within_module_init()'s behaviour changes when module unloading is disabled, both architecture would need to take this into account when counting the PLTs. Today neither architecture does this, meaning when module unloading is disabled there are insufficient PLTs in the init section to load some modules, resulting in warnings: | WARNING: CPU: 2 PID: 51 at arch/arm64/kernel/module-plts.c:99 module_emit_plt_entry+0x184/0x1cc | Modules linked in: crct10dif_common | CPU: 2 PID: 51 Comm: modprobe Not tainted 6.5.0-rc4-yocto-standard-dirty #15208 | Hardware name: QEMU KVM Virtual Machine, BIOS 0.0.0 02/06/2015 | pstate: 20400005 (nzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) | pc : module_emit_plt_entry+0x184/0x1cc | lr : module_emit_plt_entry+0x94/0x1cc | sp : ffffffc0803bba60 [...] | Call trace: | module_emit_plt_entry+0x184/0x1cc | apply_relocate_add+0x2bc/0x8e4 | load_module+0xe34/0x1bd4 | init_module_from_file+0x84/0xc0 | __arm64_sys_finit_module+0x1b8/0x27c | invoke_syscall.constprop.0+0x5c/0x104 | do_el0_svc+0x58/0x160 | el0_svc+0x38/0x110 | el0t_64_sync_handler+0xc0/0xc4 | el0t_64_sync+0x190/0x194 Instead of duplicating module_init_layout_section()s logic, expose it. Reported-by: Adam Johnston <adam.johnston(a)arm.com> Fixes: 055f23b74b20 ("module: check for exit sections in layout_sections() instead of module_init_section()") Cc: stable(a)vger.kernel.org Signed-off-by: James Morse <james.morse(a)arm.com> Signed-off-by: Luis Chamberlain <mcgrof(a)kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Zhao Wenhui <zhaowenhui8(a)huawei.com> --- include/linux/moduleloader.h | 5 +++++ kernel/module.c | 2 +- 2 files changed, 6 insertions(+), 1 deletion(-) diff --git a/include/linux/moduleloader.h b/include/linux/moduleloader.h index 2d835b7dc918..0b4541cb4c74 100644 --- a/include/linux/moduleloader.h +++ b/include/linux/moduleloader.h @@ -39,6 +39,11 @@ bool module_init_section(const char *name); */ bool module_exit_section(const char *name); +/* Describes whether within_module_init() will consider this an init section + * or not. This behaviour changes with CONFIG_MODULE_UNLOAD. + */ +bool module_init_layout_section(const char *sname); + /* * Apply the given relocation to the (simplified) ELF. Return -error * or 0. diff --git a/kernel/module.c b/kernel/module.c index 23c5349a54ea..e978debc93df 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -2301,7 +2301,7 @@ void *__symbol_get(const char *symbol) } EXPORT_SYMBOL_GPL(__symbol_get); -static bool module_init_layout_section(const char *sname) +bool module_init_layout_section(const char *sname) { #ifndef CONFIG_MODULE_UNLOAD if (module_exit_section(sname)) -- 2.34.1

2 1

[PATCH OLK-6.6 00/10] kabi: add kabi helper macros and tools
by Jialin Zhang 08 Dec '23

08 Dec '23

1. add kabi helper macros 2. add kabi tools: kabideps, kabisyms, check-kabi Li Lingfeng (1): kabi: deduplication friendly structs Xie XiuQi (5): kabi: add KABI_SIZE_ALIGN_CHECKS for more stringent kabi checks kabi: enables more stringent kabi checks kabi: add script tools to check kabi symbol kabi: add a tool to generate the kabi reference relationship kabi: add kABI reference checking tool Zheng Zengkai (4): kabi: add kabi helper macros KABI: Add CONFIG_KABI_RESERVE to control KABI padding reserve openeuler_defconfig: Enable CONFIG_KABI_RESERVE for x86 and arm64 kabi: Generalize naming of kabi helper macros Kconfig | 15 + arch/arm64/configs/openeuler_defconfig | 2 + arch/x86/configs/openeuler_defconfig | 2 + include/linux/kabi.h | 509 +++++++++++++++++++++++++ scripts/check-kabi | 147 +++++++ scripts/kabideps | 161 ++++++++ scripts/kabisyms | 141 +++++++ 7 files changed, 977 insertions(+) create mode 100644 include/linux/kabi.h create mode 100755 scripts/check-kabi create mode 100755 scripts/kabideps create mode 100755 scripts/kabisyms -- 2.25.1

2 11